Scholars who have studied global financial institutions say it’s more important for policy makers to figure out which ones are too connected to fail than ponder whether some are too big to fail. Their findings also suggest those with the potentially riskiest interconnections aren’t necessarily the biggest.
The risk of default of a large segment of the financial system depends on the network of exposure among financial institutions, the scholars say, but to date no widely accepted way of identifying the most important nodes in the financial network has been established. In a paper in Nature’s Scientific Reports, Stefano Battistan, Michelangelo Puliga, Rahul Kaushik, Paolo Tasca and Guido Caldarellui, who are economists and physicists, describe a method they have devised to figure out which banks and institutions are the most connected and therefore the most dangerous if they fail.
Josh Rothman, in the Brainiac blog at Boston.com, explains that the financial system is deeply interconnected, because banks borrow from, lend to, and own pieces of one another. In the complex financial network, the nodes are the institutions and the links are financial dependencies. The scholars created a measure called DebtRank, loosely inspired by Google’s PageRank algorithm, to analyze a huge amount of data gathered by the Federal Reserve from the $1.2 trillion FED emergency loan program to global financial institutions during 2008 -2010. The scholars found that 22 institutions, which received the bulk of the funds, formed a strongly interconnected network in which each node was systemically important. As a result, they say, a systemic default could have been triggered by small dispersed shocks. Read their paper, "DebtRank: Too Central to Fail? Financial Networks, the FED and Systemic Risk." The paper and blog show network maps of the top borrowers.
The authors note that more data would be needed to establish a fully accurate DebtRank, and much information is unavailable because of confidentiality in banking transactions. Rothman notes greater transparency in the financial system would help. The authors say their work and the DebtRank measure contributes to understanding financial networks and is also relevant to regulators and risk management practitioners. In addition, they say the work is relevant to the field of complex networks in general, and can be used to "detect systemically important nodes in any directed and weighted network.”
The Financial Stability Oversight Council, the nation’s largest financial regulatory agency created by the Dodd Frank Act, is not mentioned in the blog or the article. But it is mandated to evaluate "interconnectedness” among other characteristics such as size, scope, concentration and mix of activities in determining whether certain nonbank financial institutions could pose a threat to U.S. financial stability and therefore require enhanced supervision. Risk identifiers and measurements continue to be a work in progress.